摘要
准确预测风电场风电功率有助于减小风电出力波动对电网运行的影响。置信区间估计能够量化不确定性因素弓起的风电功率预测结果变动,向风电场运行决策者提供更多的信息。为此,对风电功率预测误差及其概率分布进行了分析,通过与正态分布、β分布拟合效果的对比,证明非参数核密度估计对风电功率预测误差概率分布具有较好的拟合效果;采用非参数核密度估计方法对风电功率预测误差进行了置信区间估计,给出了不同置信度条件下风电功率波动区间的确定方法;结合美国德克萨斯州某风电场的风电数据绘制了风电场在不同置信度条件下实际出力的波动区间,计算了对应的预测区间覆盖率,验证了基于非参数核密度估计的风电功率预测误差置信区间估计效果的有效性。
The accurate prediction of wind power can help reduce the influence of wind power output fluctuation on power grids. Confidence interval estimation can quantitate the variety of wind power prediction results and provide more information to decisionmakers. Wind power prediction error(WPPE) and its probability distribution is analyzed firstly, and non-parametric kernel density estimation is proved to have a good effect in fitting the probability distribution of WPPE compared with normal distribution and beta distribution. The non-parametric kernel density estimation is then adopted to conduct the confidence interval estimation of WPPE, and the method for determining the fluctuation intervals of the wind power under different confidence degrees is given. Based on the wind data from a Texas wind farm, the fluctuation intervals of the real wind power output and the corresponding prediction interval coverage probability(PICP) is obtained according to different confidence degrees, which verifies the effect of the proposed confidence interval estimation method.
出处
《陕西电力》
2017年第2期21-25,共5页
Shanxi Electric Power
基金
国家科技支撑计划课题资助项目(2015BAA01B00)